[clang] [llvm] [mlir] [AArch64][SME] Improve codegen for aarch64.sme.cnts* when not in streaming mode (PR #154761)

Kerry McLaughlin via llvm-commits llvm-commits at lists.llvm.org
Wed Sep 10 06:44:20 PDT 2025


================
@@ -4304,9 +4304,10 @@ Value *CodeGenFunction::EmitSMELd1St1(const SVETypeFlags &TypeFlags,
   // size in bytes.
   if (Ops.size() == 5) {
     Function *StreamingVectorLength =
-        CGM.getIntrinsic(Intrinsic::aarch64_sme_cntsb);
+        CGM.getIntrinsic(Intrinsic::aarch64_sme_cntsd);
     llvm::Value *StreamingVectorLengthCall =
-        Builder.CreateCall(StreamingVectorLength);
+        Builder.CreateMul(Builder.CreateCall(StreamingVectorLength),
+                          llvm::ConstantInt::get(Int64Ty, 8), "svl");
----------------
kmclaughlin-arm wrote:

There is not, I've added the flags here too.

https://github.com/llvm/llvm-project/pull/154761


More information about the llvm-commits mailing list